Image analysis – Applications – Personnel identification
Reexamination Certificate
2001-10-01
2004-06-29
Johnson, Timothy M. (Department: 2625)
Image analysis
Applications
Personnel identification
C382S204000, C340S005830
Reexamination Certificate
active
06757411
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to fingerprint authentication and, more specifically, to a method and system for mapping a fingerprint image into a code such that the code resulting from fingerprint images taken from a same donor finger under varying environmental conditions exhibits invariance.
BACKGROUND OF THE INVENTION
In order for fingerprint recognition to gain widespread acceptance as a reliable live-scan authentication tool, two major hurdles need to be overcome. The first involves obtaining a clear fingerprint image from a finger that may exhibit a variety of skin properties and the second involves obtaining high authentication accuracy based on fingerprint images.
Optical live-scan fingerprint scanners normally employ the mechanism of total internal reflection in a dark field, where the angle of observation from a camera is set at a so-called “glass/air critical angle”. In a two-dimensional picture, a transparent glass or plastic prism can be represented by an inverted, truncated triangle with the longer of two parallel surfaces being the platen used for making impressions of a live finger. Light illumination can come though the cut-off apex from below (i.e., the shorter parallel surface) or, alternatively, through one of the two nonparallel faces, with the camera pointing at the other nonparallel face.
Optical configurations such as the one described herein above are more rugged and produce images with higher resolution than most non-optical sensing methods such as capacitive, electric field and contact sensing. However, problems associated with so-called “keystone distortion” and problems due to fingertips with dry skin have been identified as obstacles for its usefulness as a biometric device.
Keystone distortion can be minimized by the theory of tilted planes, in combination with a software equalizer algorithm performed on an expanded image bit map to trade image size for linearity. An optical solution is to incorporate geometric equalization to compensate for the differences in distance travelled by light rays, by introducing a complementary prism between the platen prism and the camera.
The so-called “dry finger” problem is a more serious issue. In order to understand this problem, one can begin by examining what makes a print. Looking from a plan view, the skin surface of a fingertip consists of a number of curved ridges forming arches, loops and whorls, as well as composite and accidental flow patterns. Viewed in elevation under magnification, a segment of a length of ridge structure is made up of minute elevations (papillae) of uneven heights lining up in a row. This is why an inked impression of fingerprint shows broken lines and is subject to smudging and smearing of ink.
Pores exist only on ridges, and they are the opening of sweat glands under the epidermis. People who do not have dry skin problems usually have a deposit of sweat beads sitting on top of the pores. These liquid beads of sweat play a very critical role in latent print detection and render optical impression feasible. However, people with dry skin do not have as many sweat beads, which makes image taking difficult. The reason for this is now explained.
An image shows because there is intensity contrast amongst information-carrying pixels in the incident photons. If one considers the geometry of the prism as an inverted triangle, one can set up a dark field for observation purposes when the optical axis of the camera looking upwards from one side is aligned at the glass/air critical angle. Proper light illumination from underneath the prism will traverse through the transparent material of the prism and exit therethrough. Most of the ambient light surrounding the platen surface cannot reach the camera because reverse geometry of Snell's law dictates that most of the photon energy entering the body of the prism from above will be confined to below the glass/air critical angle. Dark field photography can thus be made sensitive under a noisy environment.
U.S. Pat. No. 3,527,535 to John N. Monroe, uses the term “diffusely reflected light” to describe the mechanism whereby the points where the ridges touch will appear white (light) and the background and spaces between ridges will appear black (no light). This refers to a live-scan process in which a finger touches the prism platen to form an interface for optical fingerprint impression taking.
U.S. Pat. No. 5,233,404, issued on Aug. 3, 1993 to James H. Lougheed and Lam Ko Chau, suggested an improved mechanism for contrast enhancement to Monroe's dark field principle. The term “dispersed light” is used in their patent to account for the manifestation of ridges in contact with the prism platen as bright images observed at above the critical angle.
The term “absorbed/dispersed light” is also employed in U.S. Pat. No. 5,416,573, issued on May 16, 1995 to Thomas F. Sartor, Jr., to describe internally reflected light where the finger ridges contact the prism surface. This patent discloses making the observation angle of the detector sufficiently large so that the fingerprint image will be free of artifacts attributable to moisture.
However, U.S. Pat. No. 5,416,573 does not explain why such a diffusion or dispersion phenomenon is not observed by the detector at the same observation angle when a flat object is in contact with the prism platen. For example, if the angle of reflectance described in U.S. Pat. No. 5,416,573 has a range large enough to cover the angle of observation, then a two-dimensional image of a three-dimensional object having a flat surface at contact should appear in the dark field. However, this is not the case, as long as the object/platen boundary region is completely flat and air-tight. Further explanation becomes necessary when a clear and legible fingerprint ridge pattern of a reproducible nature is desired, which is the case when live authentication is being considered.
If one carves out a grid pattern on the flat surface of a solid object and places this carved flat surface in contact with the platen of a dark field scanner, the air-filled grid pattern would appear dark and the solid contact regions bright. For good contrast, it is necessary that rows and columns of air trapped in cavities be tight. What seems to happen to the camera mounted above the critical angle is the formation of contrast due to the following mechanism. Take the case where source light beam enters one of the two non-parallel surfaces of a glass prism. If the object is totally flat and devoid of air pockets, some light will be reflected and enter the camera aperture. The important phenomenon is that the reflected beam of light is homogeneous and uniform, as the solid/glass boundary layer is homogeneous and uniform. Therefore, no contrast is noticeable. With air pockets trapped in the flat object surface, the solid wall forming the air volume bounced the photon energy back in all directions, including a horizontal direction parallel to the platen surface. At this air/glass boundary layer, the reflected horizontal light enters the glass medium from an air medium, and will be confined after refraction through the optical path within the glass prism to an angle equal to or below the critical angle according to Snell's law of total internal reflection in reverse. Consequently, less photon energy from the air/glass layer will enter the camera than that from the solid/glass layer due to differences in reflector geometry. It becomes clear that the degree of darkness in the dark image field is only relative, and it is the contrast ratio that will largely define the picture quality.
Live-scan devices make use of the difference in heights between a ridge and a valley of the fingertip skin surface. A ridge is typically tens of microns (one millionth of a meter) higher than a valley on a free standing up-turned finger. Upon turning the fingertip 180 degrees and being pressed against a platen, the difference in heights is reduced somewhat depending on the amount of applied pressure, but is still ten or more times larger than the thi
Bayat Ali
Johnson Timothy M.
Liska Biometry Inc.
LandOfFree
Method and system for fingerprint encoding and authentication does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for fingerprint encoding and authentication, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for fingerprint encoding and authentication will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3346376